Electrophysiological and Psychological Correlates of Cognition in Adult Homozygous Sickle Cell Disease Patients

Abstract

Background

Neurocognitive dysfunction is the most common neurological complication seen in homozygous sickle cell disease (SCD) patients, and it appears to be associated with the severity of anaemia, indicating hypoxic brain injury as a causative factor, but the studies are limited to the paediatric population.

Purpose

We aimed to investigate the neurocognitive profile in adult homozygous SCD patients by subjective (Montreal Cognitive Assessment [MoCA] questionnaire) as well as objective (P300 event-related potentials [ERPs]) cognitive tests.

Method

Thirty adult homozygous SCD patients and 30 age-matched healthy controls were recruited. The MoCA questionnaire was employed for subjective cognitive status. ERP was recorded by using a modified 3-stimulus auditory oddball paradigm, which contains frequent, rare and novel stimuli. Amplitudes and latencies of corresponding original and difference ERP components were measured and analysed independently. Source analysis was done using sLORETA for P3a and P3b ERP components and compared between healthy controls and SCD patients.

Results

We found a significant reduction in total MoCA scores as well as their domains in SCD patients as compared to controls. For ERP, we found reduced amplitudes and increased latencies of P300 in SCD patients. This phenomenon was further validated in analysis of difference P3 components (P3dT i.e. target minus standard and P3dN i.e. novel minus standard). Additionally, we also found negative correlation between P3a/P3dN latency and attention domain of MoCA. Further, we found a diminished source activity for P3a and P3b in SCD as compared to control subjects.

Conclusion

These results indicate impaired neurocognitive abilities in cognitive domains of adult homozygous SCD patients like attention, executive functions and working memory, and reduced source activities of P300 components which may be due to chronic cerebral hypoxia in these patients.

Keywords

Sickle cell disease cognition event-related potential P3a P3b sLORETA source localisation

Introduction

Sickle cell disease (SCD) is a prevalent genetic disorder affecting populations globally, particularly those from Sub-Saharan Africa, South America and the Mediterranean origin.¹ It is characterised by a mutation in the gene responsible for encoding the β subunit of haemoglobin, resulting in the substitution of glutamic acid with valine at the sixth position of the β globin chain. This mutation can occur heterozygously (HbAS) or homozygously (HbSS), leading to carriers of the defective gene or individuals with the disease, respectively.² In low oxygen conditions, sickled haemoglobin forms polymers within red blood cells, causing stiffening of the cell membrane and leading to haemolysis. This process results in complications collectively known as vaso-occlusive crisis, affecting various aspects of the vasculature.^{3, 4} Neurological complications, including ischaemic stroke, silent cerebral infarction (SCI), headaches and cognitive impairment, are prevalent in SCD, involving deficiencies in working memory, verbal learning, visuomotor skills, intellectual functioning, executive functions, language and attention.^{5, 6} Approximately one-third of children with SCD experience SCIs during childhood, with over half affected by young adulthood.⁷ Mechanisms behind these neurocognitive impairments include chronic anaemia that leads to hypoxic brain injury. Frequent pain episodes, along with their associated sleep disruption and psychological burden, may further impair cognitive abilities in these patients. Also, recurrent small-vessel occlusion and microinfarcts further disrupt white matter integrity and long-range connectivity important for attention, processing speed and executive function, while persistent inflammation, oxidative stress and endothelial injury exacerbate blood–brain-barrier dysfunction. Further, impaired cerebrovascular reactivity and autoregulation limit the brain’s ability to meet metabolic demands, thereby increasing vulnerability to ischaemia, producing cognitive vulnerabilities even without overt infarction.^5–7

Assessment methods for diagnosing cognitive impairment include formal psychological tests like the Montreal Cognitive Assessment (MoCA), which has high sensitivity and specificity for detecting mild cognitive impairment (MCI). Studies utilising MoCA have identified cognitive impairment in SCD patients.^{8, 9} However, neuroimaging techniques, particularly event-related potentials (ERPs) acquired through electroencephalography (EEG), are always considered superior for objective screening, diagnosing, and treating brain damage in individuals with SCD. ERPs are small electrical signals in the brain recorded via EEG in response to sensory, cognitive or motor events. Among ERPs, the P300 wave is extensively studied and plays a vital role in understanding attention, memory and cognitive control.¹⁰ It is a positive deflection observed across the scalp, with the highest amplitudes over midline central and parietal electrodes, occurring approximately 300 milliseconds (ms) after the onset of infrequent target stimuli in a standard 2-stimulus oddball paradigm.¹¹ Whereas, the modified 3-stimulus oddball paradigm introduces a rare distractor stimulus alongside target and standard stimuli, eliciting two distinct P3 responses: P3a, triggered by rare non-target stimuli reflecting involuntary attention shifts and P3b, triggered by target stimuli, indicates sustained focus on task-relevant information. More specifically, the amplitude of the P3b has been considered as an index of voluntary attention, with larger amplitudes indicating a greater allocation of attentional resources, while the peak latency of the P3b is thought to be related to the time required for stimulus evaluation.^12–15 These two components, P3a and P3b, provide valuable information about the allocation of attentional resources, cognitive processing speed, and the revision of mental representations.^{13, 15} Abnormalities in P300 ERPs, including alterations in amplitude, latency and scalp distribution, are consistently observed in children with SCD, reflecting cognitive processing and neural functioning deficits associated with the condition.^16–18

However, the true magnitude of cognitive burden of SCD in adult population is still unknown with studies done mostly on paediatric age group.¹⁹ Also, people with SCD are at risk of cognitive deficits across all domains throughout their lifespan.²⁰ Besides, scarce literature is available utilising the objective validation of the reported cognitive changes in adult SCD patients using techniques like EEG and ERP recording. In addition, EEG-based imaging techniques enable investigation of brain functioning aspects that standard imaging techniques do not and application of this in SCD patients is still awaited.

Thus, in this study, subjective and objective cognition in adult SCD patients and healthy subjects were assessed using the MoCA questionnaire and P3 ERPs, respectively. A modified auditory oddball paradigm was utilised to record P3a and P3b components of ERPs. The neural sources of P3a and P3b were identified using standardised low-resolution brain electromagnetic tomography (sLORETA), a neuroimaging method that estimates cortical neural activity distribution from EEG data with minimal localisation errors.²¹ To the best of our knowledge, this is the first study analysing both P3a and P3b neural sources in adult SCD patients, aiming to provide significant insights into cognitive processing in this population.

Methodology

Participants

This was a hospital-based case-control study in which 30 right-handed adult homozygous SCD patients and 30 age-matched healthy controls, who met the inclusion criteria, were recruited. Sample size was determined by calculating power of the study at 80%, confidence interval at 95% and significance level α = 0.05, using standard formula for case control studies, that is, n = [Z² × p × (1 – p)]/d² {where Z is 1.96, p is 3 and (considering prevalence of 3% in Chhattisgarh) and d is the tolerable sampling error, 0.05} which was equal to 45 for both the groups. Taking 15% to 20% rejection of recorded electrophysiological data during unforeseen circumstances, a minimum of 30 SCD and 30 control subjects were finally included for the study.²² Initial screening of SCD patients was performed at the department of General Medicine OPD. Pregnant or breastfeeding women and SCD patients with chronic inflammatory conditions such as SLE, rheumatoid arthritis, or any other infectious process were excluded from the study. Healthy control volunteers were included based on the absence of personal and/or family history of major medical, psychiatric, or substance-related disorders. All participants provided informed consent for their participation, which was obtained with the approval of the institute’s ethics committee.

Materials and Methods

Assessment of Subjective Cognition by MoCA Questionnaire

The MoCA questionnaire consisted of a 30-point test which was administered to all the participants.²³ Each participant filled the MoCA questionnaire according to the instructions which measured cognition across all domains, that is, attention and concentration, executive functions, memory, language, visuo-constructional skills, conceptual thinking, calculation and orientation. Scores for each domain were then calculated followed by totalling of all the scores to obtain the total MoCA score for each subject.

EEG Data Acquisition and Analysis

Continuous EEG data was collected utilising a 64-channel saline-based electrode cap from the Brain Electro Scan System (BESS) (Axxonet System Technologies Pvt. Ltd.), arranged following the extended International 10–20 System, with the ground at AFz and reference at Cz. The data were recorded utilising the BESS acquisition system and software, with impedances kept below 50 kΩ. The EEG data were recorded at a sampling rate of 1,000 Hz, and later the continuous EEG data were processed offline using the same BESS software. The data were re-referenced to a common average of all electrodes and filtered using a band-pass filter ranging from 0.5 to 35 Hz, along with a notch-filter set at 50 Hz. Subsequently, the continuous data were visually examined, and any noisy channels were interpolated based on an average of the surrounding electrodes to eliminate channel-level artefacts.

These continuous EEG data were then epoched from 200 ms prior to stimulus presentation (Frequent, Target and Novel) to 800 ms after stimulus onset, as P300 arises approximately between 300–800 ms after stimulation. A baseline correction of 200 ms pre-stimulus to 0 ms was applied to all epochs. The epochs were then examined and rejected as appropriate based on visual inspection. The remaining cleaned epochs were averaged, and peak amplitude and latency were computed at midline electrodes (Fz, FCz, CPz and Pz). The ERP components of interest were identified as the most prominent positive peaks occurring within specific time intervals: 250–350 ms for the P3a and 350–550 ms for the P3b. These time frames were chosen based on a visual examination of the grand-averaged ERPs for each condition. The peak amplitude was determined in relation to the pre-stimulus baseline, while the peak latency was measured from the time of stimulus onset.

3-stimulus Auditory Oddball Paradigm

Each subject was presented with three types of auditory stimuli consisting of a frequent tone (1,000 Hz), a target tone (500 Hz), and a novel tone (white noise). Each tone was presented for a duration of 50 ms followed by an inter-stimulus interval of 1,200 ms with a 20% variance. The subjects were instructed to close their eyes during the presentation of the protocol and press a designated key on the numpad when they hear the target tone only and refrain from pressing the key on frequent and white noise. A 1,000 ms wait time is given to the subject within which if no response is given by them, the next auditory tone is automatically presented.

Each block maintained an 80%–20% ratio between the presentation of frequent, target and distractor stimuli, with 10% each for rare and distractor stimuli. Consequently, the total number of presentations amounted to 400 (320 frequent, 40 rare and 40 novel) stimuli. The latency (time in ms from the onset of stimulus to the occurrence of the positive peak) and amplitude (distance from the baseline to the highest point of the positive peak in µV) of P300 (P3a & P3b) during the auditory 3-stimulus oddball paradigm were recorded and analysed. To consistently observe and more accurately evaluate the target and novel effects, difference waveforms were generated by subtracting the ERPs elicited by standard stimuli from those elicited by target and novel stimuli, respectively. The relevant components in the difference waveforms were then further analysed.²⁴ Besides, reaction time (msec) was also recorded for target stimuli.

Source Localisation

We used Brainstorm software for source estimation of P3a and P3b components, which is an open-source tool for analysing brain recordings such as EEG data. Brainstorm shares a complete set of easy-to-use tools with the scientific community using MEG/EEG as an experimental technique.²⁵ We created a protocol in which we used ICBM152 MRI template with 15,002 vertices and 29,984 faces.²⁶ Artefact-free ERP files were imported for each subject, and channel numbers were defined. Spherical head models were computed and source estimation was performed using the Minimum-norm imaging (MNI) and sLORETA method for each stimulus and subject. Cortical surface data were visualised using the Desikan-Killiany (DK) atlas, which contains 34 cortical regions of interest in each hemisphere, widely used for neuroimaging analyses.^{27, 28}

Statistical Analysis

The data were analysed using statistical software program (SPSS version 20, IBM, IL, USA). The comparisons of two groups in demographic features, anthropometric variables, psychological variables and ERP components were analysed by Student’s t-test for independent samples. Quantitative data were expressed as Mean ± Standard Deviation (SD). Repeated-measures analysis of variance (ANOVA) was then used to analyse the amplitude and latency of the original ERP components, with stimuli (Frequent, Target and Novel) and locations (Fz, FCz, CPz and Pz) being within-subject factors, while the groups (SCD patients vs controls) taken as a between-subject factor.²⁴ More specifically, P3 elicited by the target was defined as P3b, and P3 component elicited by distractor stimuli was defined as P3a. Midline electrodes were chosen due to the fact that P3 reaches its highest amplitude at midline sites.²⁹ For difference in ERP components (P3dT: target minus standard; P3dN: novel minus standard), repeated measures ANOVA was performed with locations (Fz, FCz, CPz and Pz) as a within-subject factor, while with group (SCD patients vs healthy controls) as a between-subject factor.²⁴ For analyses where the assumption of sphericity was not met, Greenhouse–Geisser correction was used and adjusted degrees of freedom were reported. Further post hoc analysis using Bonferroni correction was conducted if necessary. To evaluate the relationship between cognitive performance (MoCA) and electrophysiological parameters, Spearman’s correlation coefficients were calculated. Results with p < .05 were considered to be significant.

The sources of neural activity were estimated using sLORETA in the Brainstorm software. Statistical significance of differences for sLORETA elicited by the target and non-target (novel) stimuli were assessed with permutation-based voxel-by-voxel paired t-tests (within-group differences) and independent t-tests (between-group differences) based on the theory of randomisation, 5,000 random permutations were performed.^{30, 31} The Desikan-Killiany (DK) atlas was used to see the activation in cortical areas. Brodmann areas and MNI coordinates were only extracted for statistically significant sources (p < .05).

Results

Sample Characteristics

Among the 30 participants in the SCD group, 18 were males and 12 were females, while the control group comprised of 17 males and 13 females. In this hospital-based study, adult patients who presented for treatment were recruited irrespective of time of diagnosis, as this is an inherited disease. All SCD patients were on hydroxyurea therapy with no history of any cerebral infarcts. All the participants in our study were between the ages of 18–40 years (SCD: 25.76 ± 5.09; control: 26.83 ± 5.68). SCD patients had significantly lower BMI compared to control subjects (17.91 ± 2.46 vs 22.48 ± 4.11; p < .001). We also analysed different blood indices between these two groups and SCD patients had significantly low Hb (9.54 ± 1.43 vs 13.69 ± 2.04), HCT (28.70 ± 4.13 vs 42.63 ± 5.63), RBC count (3.59 ± 0.79 vs 5.04 ± 0.82) (p < .001) than control subjects, respectively.

Assessment of Cognition

1. MoCA Questionnaire:

Subjective cognition, indicated by total MoCA scores, reveals that SCD individuals exhibited significantly low total MoCA scores compared to control subjects (20.43 ± 3.09 vs 26.16 ± 2.67; p < .001). Components of MoCA, namely executive function, naming, attention, language, abstract and delayed recall also shows significant difference (Figure 1).

Figure 1.

Note: **p < .01, ***p < .001. SCD: Sickle cell disease.

2. Electrophysiological Findings:

a. Characterisation of Original ERP

Control subjects showed higher P3a amplitude than SCD patients at Fz (2.23 ± 1.62 µV vs 2.0 ± 1.36 µV) and FCz (2.55 ± 2.1 µV vs 2.0 ± 1.5 µV), whereas P3b amplitude are similar in both groups at CPz (2.56 ± 1.6 µV vs 2.46 ± 1.62 µV) and Pz (3.21 ± 2.03 µV vs 3.25 ± 1.78 µV). Similarly, controls showed shorter P3a latency than SCD patients at Fz (318.50 ± 32.73 ms vs 321.56 ± 24.65 ms) and FCz (316.10 ± 41.89 ms vs 318.63 ± 26.95 ms) whereas latency of P3b for controls at CPz site was shorter than SCD (355.60 ± 33.35 ms vs 358.93 ± 40.89 ms) but, longer for controls than SCD at Pz site (364.06 ± 40.74ms vs 361.46 ± 42.61ms), though they were not statistically significant. Figures 2 and 3 represent the grand average waveforms and topographic maps of the original ERP components.

Figure 2.

The Grand Average Waveforms of ERP in SCD Patients and Healthy Subjects were Analysed and Drawn. Four Electrodes Include: Fz, FCz, CPz and Pz. Amplitude of P3a (at Fz and FCz) and P3b (CPz and Pz) for Control was Larger than SCD Patients, Whereas Latency of P3a (at Fz and FCz) and P3b (CPz and Pz) for Control was Shorter as Compared to SCD Thereby Indicating Impaired Attentional Allocation, Cognitive Processing Efficiency, and Delayed Stimulus Evaluation in SCD Patients, Likely Reflecting Impaired Neurocognitive Performance in Terms of Attention, Executive Functions and Working Memory.

Figure 3.

Topographical Voltage Distribution of P3 in Healthy Subjects and SCD Patients. More Frontal Distribution of P3a for Control is Seen as Compared to SCD, Whereas P3b Distributed Parietally for Both Controls and SCD Patients Thereby Indicating Preserved Target Evaluation Processes Across Groups But Impaired Frontal Attentional Engagement and Orienting Responses in SCD Patients.

ANOVA for P3 amplitudes revealed that there was a significant main effect of stimulus [F (2,116) = 102.07, p < .001, partial η² = 0.638], with post hoc comparisons showing largest P3a elicited by novel stimuli (2.54 ± 1.77 µV). We also found significant main effect of location [F (2.17, 125.92) = 13.68, p < .001, partial η² = 0.191] which was qualified by the two-way interaction of stimulus × location [F (3.51, 203.57) = 3.24, p = .01, partial η² = 0.052], wherein P3a elicited by novel stimuli was the largest (2.32 ± 1.82 µV) at FCz site while P3b elicited by target stimuli was largest (3.23 ± 1.89 µV) at Pz site.

In terms of P3 latency, SCD patients displayed a longer P3 latency (352.17 ± 39.70 ms) than healthy controls (336.70 ± 43.66 ms, [F (1, 58) = 9.422, p = .003, partial η² = 0.140)], and also a remarkable stimulus effect [F (2, 116) = 11.44, p < .000, partial η² = 0.164)]. For the stimulus effect, post hoc comparisons showed longest duration for P3b (elicited by target stimuli) (351.46 ± 41.57 ms). There was a remarkable group × stimulus interaction [F (2, 116) = 7.52, p = .001, partial η² = 0.114)] wherein shorter P3a latency and longer P3b latency in SCD group was indicated as compared to controls (323.54 ± 35.92 ms vs 339.09 ± 44.30 ms for P3a), (352.97 ± 41.88 ms vs 349.95 ± 41.68 ms for P3b), although the results are not statistically significant. We also found significant location effect [F (2.10, 121.82) = 7.26, p = .001, partial η² = 0.111)] and post hoc analysis showed that FCz site has shortest latency (336.52 ± 42.14 ms) among all four electrode sites. There was also significant stimulus × location effect [F (3.65, 212.18) = 4.08, p = .004, partial η² = 0.066)] and post hoc analysis showed shortest latency for Novel stimulus at FCz site (317.36 ± 34.94 ms). Lastly, we found significant effect of group × stimulus × location [F (6, 348) = 2.23, p = .03, partial η² = 0.037)] and post hoc analysis showed shortest latency in controls and SCD patients for Novel stimulus at FCz site (316.10 ± 41.89 ms vs 318.63 ± 26.95 ms) (Table 1). We also found higher response time, though statistically nonsignificant, for P3 in SCD group than that of control subjects (525.62 ± 93.17 ms vs 472.31 ± 93.17ms).

Table 1.

Summary of Repeat Measures ANOVA for Comparison of Study Groups, Type of Stimuli and Recording Locations on Amplitude and Latency of P300, Indicating Significant Interaction of Stimulus and Location on Amplitude, and Significant Interaction of Group, Stimulus and Location on Latency.

Variables	Source	df	F Value	p Value	Partial Eta Squared
Amplitude	Groups	1,58	0.02	.898	0
	Stimulus	2,116	102.07	< 0.001***	0.638
	Location	2.17, 125.92	13.68	< 0.001***	0.191
	Groups × Stimulus	2,116	0.09	.91	0.002
	Groups × Location	2.17, 125.92	0.89	.418	0.015
	Location × Stimulus	3.51, 203.57	3.24	.01*	0.052
	Groups × Location × Stimulus	3.51, 203.57	0.61	.645	0.01
Latency	Groups	1,58	9.42	.003**	0.139
	Stimulus	2,116	11.44	< 0.001***	0.164
	Location	2.1, 121.82	7.26	< 0.001***	0.111
	Groups × Stimulus	2,116	7.52	< 0.001***	0.114
	Groups × Location	2.1, 121.82	1.79	.168	0.03
	Location × Stimulus	3.658, 212.18	4.09	.004**	0.065
	Groups × Location × Stimulus	3.658, 212.18	2.24	.072	0.0371

Notes: df: Degree of freedom.

*p < .05, **p < .01, ***p < .001.

b. Characterisation of Differences of ERP

For a better evaluation of the effects of target and novel stimuli, we measured the P3dN (novel minus frequent) and P3dT (target minus frequent) components also. Control subjects showed higher P3dN amplitude than SCD patients at Fz (2.81 ± 1.81 µV vs 2.17 ± 1.56 µV) and FCz (3.42 ± 2.48 µV vs 2.43 ± 1.80 µV), similarly P3dT amplitude for control is higher than SCD at CPz (2.50 ± 1.70 µV vs 2.39 ± 1.69 µV) but lower than SCD at Pz (2.62 ± 2.13 µV vs 3.13 ± 2.16 µV). Similarly, controls showed shorter P3dN latency than SCD patients at Fz (310.46 ± 29.40 ms vs 319.23 ± 24.28 ms) and FCz (306.30 ± 30.64 ms vs 317.70 ± 27.13 ms), whereas latency of P3dT for controls was shorter than SCD at CPz site (348.53 ± 30.47 ms vs 359.36 ± 39.10 ms) and at Pz site (355.63 ± 33.47 ms vs 364.40 ± 38.51 ms), though they were not statistically significant. Figure 4 depicts the plotted waveforms for the difference of P3 between both groups. Figure 5 depicts the topographical voltage distribution of the difference of ERPs.

Figure 4.

The Grand Average Waveforms of Difference ERP Components in SCD Patients and Healthy Subjects. Amplitude of P3dN (at Fz and FCz) and P3dT (CPz and Pz) for Control was Larger than SCD Patients, Whereas Latency of P3dN (at Fz and FCz) and P3dT (CPz and Pz) for Control was Shorter as Compared to SCD. This Indicates Impaired Cognitive Processing Efficiency in SCD Patients, Characterised by Reduced Allocation of Attentional and Working Memory Resources, Along with Delayed Neural Responses to Task-relevant and Distractor Stimuli.

Figure 5.

Topographical Voltage Distribution in Two Groups. P3dT (Target Minus Standard) and P3dN (Novel Minus Standard) Represent P3 Target and Novel Effects. More Frontal Distribution of P3dN for Control is seen as Compared to SCD, Whereas P3dT Distributed Parietally for Both Controls and SCD Patients. This Indicates Impaired Frontal Attentional Engagement and Novelty Detection Mechanisms in SCD Patients, While the Parietal Processing of Task-relevant Target Stimuli Remains Relatively Preserved.

We did ANOVA for difference ERPs also. The amplitude analysis of P3dN and P3dT revealed significant location effects for P3dT [F (1.97, 114.32) = 16.48, p < .001, partial η² = 0.221)] with post hoc analysis showing highest amplitude of P3dT at Pz (2.88 ± 2.14 µV). Significant interaction of group × locations for P3dN was also observed [F (1.93, 112.11) = 6.18, p = .003, partial η² = 0.096)] where post hoc analysis revealed highest amplitude of P3dN for SCD group at CPz site (2.67 ± 1.99 µV) and at FCz site for control (3.38 ± 2.45 µV) (Table 2).

The latency analysis of P3dN and P3dT showed significant main effect of location [F (1.79, 104.33) = 8.78, p < .001, partial η² = 0.132 for P3dN]; [F (2.09, 121.66) = 4.43, p = .013, partial η² = 0.07 for P3dT)] and post hoc analysis showed shorter latency at FCz site for both P3dN (312 ± 29.26 ms) and P3dT (343.55 ± 39.54 ms) (Table 2).

Table 2.

Summary of Repeated Measures ANOVA for Comparison of Study Groups and Recording Locations on Amplitude & Latency of P3dN and P3dT.

Variables	Source	df	F Value	p Value	Partial Eta Squared
P3dN Amplitude	Groups	1, 58	0.03	.87	0
	Location	1.93, 112.11	2.73	.07	0.045
	Groups × Location	1.93, 112.11	6.18	.003**	0.096
P3dN Latency	Groups	1, 58	0.69	.41	0.994
	Location	1.79, 104.33	8.78	< 0.001***	0.132
	Groups × Location	1.79, 104.33	0.59	.54	0.01
P3dT Amplitude	Groups	1, 58	0.04	.83	0
	Location	1.97, 114.32	16.45	< 0.001***	0.221
	Groups × Location	1.97, 114.32	0.71	.49	0.012
P3dT Latency	Groups	1, 58	1.09	.3	0.018
	Location	2.09, 121.66	4.44	.012*	0.071
	Groups × Location	2.09, 121.66	0.16	.86	0.002

Notes: Findings indicate significant interaction of groups and location on P3dN amplitude, and significant effect of location on P3dN latency, P3dT amplitude and P3dT latency.

df: Degree of freedom.

*p < .05, **p < .01, ***p < .001.

3. Correlations Between Subjective and Objective Components of Cognition

We correlated the amplitude and latency of P3a (at Fz and FCz), P3b (at CPz and Pz), P3dN (at Fz and FCz), and P3dT (at CPZ and Pz) elicited by distractor and target stimulus with the MoCA scores to explore in depth relationship of these ERP components with subjective cognitive state (Table 3). Negative correlation of P3b and P3dT amplitude with abstract component of MoCA, and P3dN amplitude with orientation component of MoCA were evident. Further, negative correlation of P3a and P3dN latency with attention domain and total MoCA scores was observed.

Table 3.

Correlation of ERP Parameters with MoCA Scores in SCD Patients.

ERP Components/MoCA Components	Executive	Naming	Attention	Language	Abstract	Delayed Recall	Orientation	MoCA Total
P3a Amplitude (Fz)	NS	NS	NS	NS	NS	NS	NS	NS
P3a Amplitude (FCz)	NS	NS	NS	NS	NS	NS	NS	NS
P3b Amplitude (CPz)	NS	NS	NS	NS	NS	NS	NS	NS
P3b Amplitude (Pz)	NS	NS	NS	NS	−.396* (0.03)	NS	NS	NS
P3dN Amplitude (Fz)	NS	NS	NS	NS	NS	NS	NS	NS
P3dN Amplitude (FCz)	NS	NS	NS	NS	NS	NS	−.373* (0.04)	NS
P3dT Amplitude (CPz)	NS	NS	NS	NS	NS	NS	NS	NS
P3dT Amplitude (Pz)	NS	NS	NS	NS	−.366* (0.04)	NS	NS	NS
P3a Latency (Fz)	NS	NS	−.335* (0.04)	NS	NS	NS	NS	NS
P3a Latency (FCz)	NS	NS	−.370* (0.04)	NS	NS	NS	NS	−.371* (0.04)
P3b Latency (CPz)	NS	NS	NS	NS	NS	NS	NS	NS
P3b Latency (Pz)	NS	NS	NS	NS	NS	NS	NS	NS
P3dN Latency (Fz)	NS	NS	−.383* (0.03)	NS	NS	NS	NS	NS
P3dN Latency (FCz)	NS	NS	−.363* (0.04)	NS	NS	NS	NS	NS
P3dT Latency (CPz)	NS	NS	NS	NS	NS	NS	NS	NS
P3dT Latency (Pz)	NS	NS	NS	NS	NS	NS	NS	NS

Notes: The findings indicate significant negative correlation of P3dN amplitude with orientation component of MoCA, and P3b and P3dT amplitude with abstraction domain of MoCA. For latency, it indicates significant negative correlation P3a and P3dN latency with attention domain and total MoCA scores.

*p < .05.

NS: Not significant.

d. Source Localisation

i. Comparison of P3a and P3b Within Control and SCD Patients

sLORETA t-test maps, for comparison among stimulus types (P3a, P3b) within control and SCD patients are depicted in Figure 6a and 6b, respectively. For each comparison, a detailed summary of the t-scores and MNI coordinates for the activated cortical regions are given in Tables 4 and 5. We analysed sLORETA images from targets and non-targets (novel) to determine which brain areas were activated differently by these stimuli. For controls, the sLORETA current source density maps for the P3a time frame showed increased bilateral activation in some brain areas in response to non-targets versus targets. Specifically, we found significant increase of P3a source activity in frontal regions (bilateral superior frontal gyri, bilateral anterior cingulate gyri, bilateral anterior portion of paracentral lobule, right frontal pole, right lateral orbitofrontal), temporal regions (right superior, middle and inferior temporal gyrus), parietal (bilateral posterior cingulate gyri, bilateral precuneus and right post central gyrus). For SCD patients, we observed significantly higher activation of P3a source in frontal regions (bilateral superior frontal gyri, bilateral anterior cingulate gyri, right lateral orbitofrontal, right frontal pole, right precentral, left inferior frontal gyrus and left mid frontal gyrus), and parietal regions (bilateral precuneus, right post central gyrus, right superior parietal, left anterior cingulate gyrus and left isthmus cingulate region) (p < .05).

For controls, the sLORETA current source density maps for the P3b showed significant increased activity in bilateral mid frontal, bilateral parietal (precuneus), inferior parietal and hippocampus. On the other hand, SCD patients showed significant increase in bilateral mid frontal, and parietal regions (right post central gyrus, right supramarginal gyrus, left inferior temporal, left superior and inferior parietal) (p < .05).

Figure 6.

sLORETA Three-dimensional Maps of Voxel-by-voxel Paired Sample t-statistics Representing Target Minus Non-target (Novel) Difference, Corresponding to the P3a (Upper Panel) and the P3b (Lower Panel) Latency Windows for Healthy Controls (a) and SCD Patients (b). The sLORETA Scales Show Negative (Blue) and Positive (Red) t-values for Which the Alpha is Significant (p < .05) After Holmes’’ Correction for Multiple Comparisons. An Increase of P3b Source Activity from P3a is Represented in Yellowish Red and Vice Versa. This Indicates Higher Source Activity for P3a as Compared to P3b for both Control and SCD Patients.

Table 4.

Differences in Brain Activation Within P3a and P3b Latency Windows in Controls.

Comparison

Area

L/R

MNI Coordinates

t Value

Comparison

Area

L/R

MNI Coordinates

t Value

Target vs Novel P3a (250–350 ms post stimulus)

Anterior cingulate

2.65

22.28

24.58

−6.31

Target vs Novel P3b (350–550 ms post stimulus)

Precuneus

13.3

−57.34

17.2

3.28

Posterior cingulate

1.68

−5.7

40.32

−5.65

Hippocampus

26.08

−36.63

3.67

4.1

Superior frontal

4.43

27.12

34.41

−5.09

Rostral middle frontal

22.51

69.75

−2.89

3.1

Paracentral

1.92

−32.35

66.8

−5.19

Precuneus

−0.14

−62.92

17.1

3.22

31,7

Precuneus

2.78

−40.72

51.97

−5.8

Rostral middle frontal

−46.76

24.63

36.6

2.69

Frontal pole

7.64

62.7

−18.89

−4.85

39,40

Inferior parietal

−47.9

−58.54

16.1

3.5

Lateral orbitofrontal

15.61

53.87

−22.24

−5.5

Caudal middle frontal

−47.69

79.8

36.5

2.65

Supramarginal

−24

26.1

−3.48

Superior temporal

70.6

−25.1

7.2

−2.59

Inferior temporal

61.8

−50.1

−22.5

−2.32

Middle temporal

70.6

−34.4

−7.3

−2.29

3,1,2

Post central

68.2

−14.3

−4.3

Anterior cingulate

−0.22

20.34

21.64

−7.2

Posterior cingulate

−0.23

−8.87

31.47

-6.4

Superior frontal

−0.3

22.34

35.38

−5.22

Paracentral

−0.98

−35.47

59.87

−5.6

31,7

Precuneus

−0.05

−48.54

−4.8

Table 5.

Differences in Brain Activation Within P3a and P3b Latency Windows in SCD Patients.

Comparison

Area

L/R

MNI Coordinates

t Value

Comparison

Area

L/R

MNI Coordinates

t Value

Target vs Novel P3a (250–350 ms post stimulus)

Anterior cingulate

1.63

24.22

27.54

−5.91

Target vs Novel P3b (350–550 ms post stimulus)

Rostral middle frontal

49.61

42.45

21.3

2.05

Superior frontal

3.57

16.59

39.29

−4.65

3,1

Postcentral

30.71

−32.2

70.32

2.43

Paracentral

11.21

−40.48

56.64

−5.2

Supramarginal

60.72

−45.2

26.18

3.6

31,7

Precuneus

6.55

−51.96

38.04

−7.4

Superior parietal

−23.5

−84.7

43.21

2.2

Lateral Orbitofrontal

9.68

56.73

-25.99

−4.34

Inferior parietal

−32.4

−85.5

36.21

2.1

Frontal pole

7.6

61.77

−17

−3.9

Inferior temporal

−51.2

−66.7

−4.66

2.4

Precentral

28.8

−25.94

50.63

−4.6

Rostral middle frontal

−40.1

54.19

−2.07

2.1

Postcentral

44.19

−29.65

65.37

−3.6

Superior parietal

43.21

−40.41

60.57

−3.5

Pars triangularis

33.22

22.86

9.53

−4.1

Anterior cingulate

−0.29

24.21

25.59

−6.4

Superior frontal

−7.96

19.38

33.33

−5.37

Posterior cingulate

−0.21

−0.97

31.42

−7.25

31,7

Precuneus

−2.93

−49.96

40.07

−5.6

Isthmus cingulate

−2.03

−49.16

35.11

−7.9

Precentral

−40.11

−15.41

34.12

−6.3

Caudal mid frontal

−33.55

8.7

28.23

−3.4

Supramarginal

−40.84

−28.96

37.77

−4.3

Rostral middle frontal

−24.57

60.01

−5.95

−3.4

Note: Tables 4 and 5. Brain regions showing decreased activation for target vs non-target stimuli within the P3a latency window (250–350 ms) and increased activation for target vs non-target stimuli within the P3b latency window (350–550 ms) at significance level (p < .05). (4) for controls and (5) for SCD patients, thereby indicating frontal areas are activated during P3a generation in both control and SCD, but, during P3b generation parietal areas (specifically precuneus) are activated in controls but not in SCD BA: Brodmann area; L: Left; R: Right.

ii. Comparison of P3a and P3b Between Control and SCD Patients

Comparison of sLORETA t-test maps for stimulus types (P3a, P3b) between control and SCD patients are depicted in Figure 7 and a detailed summary of the t-scores and MNI coordinates for the activated cortical regions are given in Table 6 respectively. For each comparison, a smaller source activity in SCD patients in comparison with controls were obtained for both components: P3a and P3b. In particular, significant reduced P3a source activity in SCD patients was mainly observed in frontal (right frontal pole, left anterior cingulate gyrus), temporal (insula and transverse temporal gyrus), and parietal (bilateral posterior cingulate gyri) (p <.05). On the other hand, P3b source activity was significantly reduced for SCD patients in frontal (left superior frontal, right middle frontal and precentral gyrus), parietal regions (bilateral inferior parietal, right post central gyrus, right superior parietal, right precuneus and supramarginal gyrus), right para hippocampal and fusiform gyrus.

Figure 7.

Differences in Brain Activation Patterns Between Controls and SCD Patients Using Voxel-by-voxel Independent t-statistics. The Scale Shows Negative (Blue) and Positive (Yellow) t-values for Which Alpha is Statistically Significant (p < .05). Positive t-values Represent Larger Source Activity in Control Group than in SCD Patients and Vice Versa for Which the Alpha is Significant After Holmes’ Correction for Multiple Comparisons. This Indicates Diminished Source Activity of P3a in Frontal and P3b in Parietal Areas (Specifically in Precuneus) in SCD Patients as Compared to Healthy Controls.

Table 6.

Results of Between-group Brain-source Generator Analyses. Stereotaxic Coordinates and Significance Level (p < .05) of Regions Show Increased Activation in Controls for P3a and P3b.

Comparison	BA	Area	L/R	MNI Coordinates			t Value	Comparison	BA	Area	L/R	MNI Coordinates			t Value
Comparison	BA	Area	L/R	X	Y	Z	t Value	Comparison	BA	Area	L/R	X	Y	Z	t Value
Control vs SCD (P3a)	10	Frontal pole	R	11.6	64.76	−14.9	4.3	Control vs SCD (P3b)	46	Rostral middle frontal	R	19.54	69.7	−1.95	3.4
	23	Posterior cingulate	R	2.62	−11.77	34.37	6.9		4	Precentral	R	20.14	−16.74	74.64	3.82
	23	Posterior cingulate	L	−0.23	−12.73	35.36	5.8		3	Postcentral	R	25.88	−27.51	67.37	3.49
	24	Anterior cingulate	L	−0.17	30.99	17.8	8.7		31	Precuneus	R	2.79	−62.91	19.13	4.87
	16	Insula	L	−42.44	−9.01	0.49	4.44		39	Inferior Parietal	R	28.65	−65.28	37.4	2.43
	41,42	Transverse temporal	L	−45.26	−17.66	8.23	5.54		28	Parahippocampal	R	22.32	−34.97	−18.7	4.24
									37	Fusiform	R	24.3	−37.91	−17.7	4.02
									8	Superior frontal	L	−6.96	54.02	1.33	3.4
									5	Superior parietal	L	−27.11	−51.56	41.84	4.45
									39	Inferior Parietal	L	−38.15	−55.68	19.13	4.43
									40	Supramarginal	L	−44.07	−25.16	18.84	3.2

Note: Table 6. Brain regions showing increased activation of P3a and P3b in controls as compared to SCD. It indicates increased activation of P3a in frontal, more specifically in cingulate areas in controls as compared to SCD patients. For P3b, it shows higher activation in frontoparietal areas in controls than SCD BA: Brodmann area; L: Left; R: Right.

Discussion

The present study attempted to assess the level of cognitive deficit in SCD patients by measuring subjective as well as objective cognition through MoCA questionnaires and P300 ERP respectively, and further to characterise P3a and P3b brain sources with comparison between them in adult SCD patients and healthy control subjects.

The MoCA emerges as a promising tool for the evaluation of cognitive impairment among adults with SCD. The total MoCA scores in SCD patients were significantly less (cutoff level >25) as compared to controls which reveals MCIs in these patients. Further analysis shows that almost all the components of MoCA, namely, executive, naming, attention, language, abstract and delayed recall ability were significantly less in SCD patients. Earlier reports also indicated that 46% to 65% of the participants failed to meet the cutoff level scores of MoCA suggestive of MCI in these patients.^{8, 9}

In the present study, we used 3-stimulus auditory oddball paradigm to evaluate P3a and P3b in terms of amplitude and latency. Low amplitudes and longer latencies of P3a and P3b at their respective locations were recorded in SCD as compared to controls, though statistically not significant possibly because of low sample size. Previous studies of P3 ERPs in children with SCD also documented reduced P3 amplitude indicating deficits in early attention modulation.³² Furthermore, another study showed delayed cognitive responses to auditory stimuli and altered neural network activation patterns, notably in areas such as the precuneus. This reduced efficiency in neural networks could account for prolonged task completion times in SCD patients.¹⁸ Also, SCD may causes variable degree of cochlear anomalies without indication of neural problems when compared to control group with prevalent impairment in both cognitive and auditory states among children with SCD.^{17, 33} We also noticed a long reaction time of P3 in SCD which corroborates with previous study in which there was noticeable increase in reaction times and decrease in the amplitudes of ERP components compared to controls.¹⁶ These observations implied that the diminished attentiveness, performance and cognitive capabilities seen in sickle cell anaemia patients are closely linked to the severity of the condition.¹⁶

Further, we correlated amplitude and latency of P3a, P3b, P3dN and P3dT with MoCA scores. The observed negative correlations between P300 subcomponents and MoCA domains in the present study offer valuable insights into the neurophysiological basis of cognitive dysfunction. Specifically, the negative correlation between P3b and P3dT amplitude with the abstract reasoning component of MoCA, suggesting diminished cognitive resource allocation or reduced efficiency in higher-order cognitive integration among SCD patients with lower abstraction abilities. Similarly, the negative association between P3dN amplitude and the orientation domain of MoCA may reflect impairments in novelty detection and contextual processing, processes mediated by frontal cortical regions. Also, the negative correlation between latency of P3a and P3dN with attention domain of MoCA and total MoCA scores which suggests slower neural processing speeds for higher-order cognitive functions specifically for attention. Taken together, these subjective and objective cognitive correlations underscore that frontal and parietal processing are definitely diminished in SCD possibly due to chronic hypoxia and cerebral vasculopathy.

Our study utilised sLORETA to determine neural correlates of P3a and P3b processing in SCD. A 3-stimulus auditory oddball paradigm was used to analyse and distinguish the brain sources of P3a and P3b, and to make comparisons between individuals with SCD and control subjects. sLORETA is a neuroimaging method that estimates cortical neural activity distribution from EEG data with minimal localisation errors.³⁰ Despite their temporal overlap, the two components exhibit dissimilarities in both their topographic characteristics and the locations of their electrical generators. We identified statistically significant differences in brain source activation between SCD patients and controls for both P3a and P3b components. A lower P3a source activation in right frontal pole (BA 10), left anterior cingulate gyrus (BA 24), transverse temporal gyrus (BA 41,42), insula (BA 16) and bilateral posterior cingulate gyrus (BA 23) was observed in SCD patients as compared to control group (Figure 7 and Table 6). As per findings, differences within the same group underscore the significance of these regions in characterising auditory 3-stimulus oddball tasks (Figure 6 and Tables 4 and 5). In this regard, superior and middle frontal, anterior and posterior cingulate areas were commonly involved in generation of auditory P3a in healthy subjects. The cerebral network responsible for orienting attention, involving the shift of attention towards novel or unexpected stimuli, aligns with the hypothesis suggesting that P3a reflects the automatic allocation of attention.^{13, 14}

On the other hand, bilateral inferior parietal, and superior, middle frontal, pre and post central gyrus shows less activation for P3b in SCD patients. Also, there was a lack of activation of precuneus area in SCD patients in our study (Figure 7 and Table 6). Similar findings in SCD children also reported lack of activation of this vital structure.¹⁸ Combining our findings with prior research, it becomes evident that frontal, cingulate and precuneus hypoactivation play a significant role in the impaired response of SCD patients in P3 paradigms. Neuropathological and functional abnormalities detected in these regions have previously been reported in SCD patients.^{18, 34} These abnormalities could potentially explain the reduced activation observed during task performance. The frontal lobe is associated with the execution of discriminatory tasks, while the cingulate cortex is believed to play a role in both the initiation and the suppression of motor responses.^{35, 36} Precuneus is involved in episodic memory, visual-spatial abilities, motor activity, coordination strategies, self-perception, executive and working memory.³⁷ The absence of activation in sLORETA analysis may be attributed to the diminished cortical thickness observed in the precuneus region in individuals with SCD. This reduction in anatomical substrate could result in a less efficient neural network, causing delays in task completion. Consequently, this could also account for the absence of correlation between P300 latencies, specifically between stimulus classification/discrimination and stimulus evaluation.^{34, 38}

Conclusion

This study which, to the best of our knowledge, is the first of its kind particularly in adult SCD patients, led to the conclusion that there was marked cognitive impairment in individuals with SCD. This decline in cognitive abilities includes various domains, including executive functions, processing speed, visual-motor coordination, vocabulary, visual memory, abstract reasoning and verbal comprehension. Various factors contribute to this impairment, including insufficient oxygen and blood flow to the brain, as well as instances of brain infarcts and strokes. Moreover, cerebral haemodynamic insufficiency, where oxygen demand surpasses supply, along with reduced oxygen saturation, further exacerbates white matter loss in SCD patients. Collectively, these results support the utility of ERPs as sensitive, non-invasive biomarkers of cognitive dysfunction in SCD.

Limitations

This study has limitations that should be acknowledged. It was conducted in a hospital-based setting and therefore, larger, community-based studies are needed to validate and extend these results across more diverse populations. Moreover, while MoCA is a widely-used cognitive screening tool, it may not capture all cognitive domains in detail.

Footnotes

Acknowledgement

The authors would like to thank all the participants for participating in the study.

Authors’ Contribution

TS: data acquisition, processing, analysis, interpretation and manuscript writing; RS and MS: conceptualisation, supervision, analysis and interpretation; PW: patient recruitment and data acquisition; AI and ST: interpretation and discussion.

Statement of Ethics

The study has been approved by the Institute Ethics Committee (Ref. No. 2053/IEC-AIIMSRPR/2021), AIIMS, Raipur.

Declaration of Conflict of Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship and/or publication of this article: Intramural Grant from AIIMS Raipur.

Patient Consent

Informed and written consent were obtained from the participants for this study.

ORCID iDs

Tarun Sahu

Meenakshi Sinha

Ramanjan Sinha

Sai Krishna Tikka

References

Serjeant

GR.

Sickle-cell disease. Lancet 1997; 350: 725–730.

Ware

, de Montalembert

, Tshilolo

, . Sickle cell disease. Lancet 2017; 15: 311-323.

Piel

, Steinberg

and Rees

DC.

Sickle cell disease. N Engl J Med 2017; 376: 1561–1573.

Kato

, Piel

, Reid

, . Sickle cell disease. Nat Rev Dis Prim 2018; 4: 1–22.

Stotesbury

, Kawadler

, Hales

, . Vascular instability and neurological morbidity in sickle cell disease: An integrative framework. Front Neurol 2019; 10: 1–21.

Sahu

, Pande

, Sinha

, . Neurocognitive changes in sickle cell disease: A comprehensive review. Ann Neurosci 2022; 29: 255–268.

Farooq

and Testai

FD.

Neurologic complications of sickle cell disease. Curr Neurol Neurosci Rep 2019; 19: 17.

Cichowitz

, Carroll

, Strouse

, . Utility of the montreal cognitive assessment as a screening test for neurocognitive dysfunction in adults with sickle cell disease. South Med J 2016; 109: 560–565.

Early

, Linton

, Bosch

, . The Montreal cognitive assessment as a cognitive screening tool in sickle cell disease: Associations with clinically significant cognitive domains. Br J Haematol 2022; 198: 382–390.

10.

Sur

and Sinha

VK.

Event-related potential: An overview. Ind Psychiatry J 2009; 18: 70–73.

11.

Picton

TW.

The P300 wave of the human event-related potential. J Clin Neurophysiol 1992; 9: 456–479.

12.

Wronka

, Kaiser

and Coenen

AML.

Neural generators of the auditory evoked potential components P3a and P3b. Acta Neurobiol Exp (Wars) 2012; 72: 51–64.

13.

Polich

Updating P300: An integrative theory of P3a and P3b. Clin Neurophysiol 2007; 118: 2128–2148.

14.

Volpe

, Mucci

, Bucci

, . The cortical generators of P3a and P3b: A LORETA study. Brain Res Bull 2007; 73: 220–230.

15.

Polich

and Criado

JR.

Neuropsychology and neuropharmacology of P3a and P3b. Int J Psychophysiol 2006; 60: 172–185.

16.

Hamon

, Seri

and Sangare

Auditory event-related potentials and reaction times during simple sensorymotor tasks in subjects with sickle cell disease and related hemoglobinopathies. Ital J Neurol Sci 1990; 11: 249–258.

17.

de Castro Silva

, Magalhães

, Toscano

, . Auditory-evoked response analysis in Brazilian patients with sickle cell disease. Int J Audiol 2010; 49: 272–276.

18.

Colombatti

, Ermani

, Rampazzo

, . Cognitive evoked potentials and neural networks are abnormal in children with sickle cell disease and not related to the degree of anaemia, pain and silent infarcts. Br J Haematol 2015; 169: 597–600.

19.

Abdi

, De Haan

and Kirkham

FJ.

Neuroimaging and cognitive function in sickle cell disease: A systematic review. Children 2023; 10(3): 532.

20.

Prussien

, Jordan

, De Baun

, . Cognitive function in sickle cell disease across domains, cerebral infarct status, and the lifespan: A meta-analysis. J Pediatr Psychol 2019; 44: 1–11.

21.

Pascual-Marqui

, Esslen

, Kochi

, . Functional imaging with low-resolution brain electromagnetic tomography (LORETA): A review. Methods Find Exp Clin Pharmacol 2002; 24 (Suppl C): 91–95.

22.

Patra

, Khodiar

, Hambleton

, . The Chhattisgarh state screening programme for the sickle cell gene: A cost-effective approach to a public health problem. J Community Genet 2015; 6: 361–368.

23.

Nasreddine

, Phillips

, Bédirian

, . The Montreal Cognitive Assessment, MoCA: A brief screening tool for mild cognitive impairment. J Am Geriatr Soc 2005; 53: 695–699.

24.

Guo

, Tian

, Xu

, . The impact of attack frequency and duration on neurocognitive processing in migraine sufferers: Evidence from event-related potentials using a modified oddball paradigm. BMC Neurol 2019; 19: 73.

25.

Tadel

, Baillet

, Mosher

, . Brainstorm: A user-friendly application for MEG/EEG analysis. Comput Intell Neurosci 2011; 2011: 879716.

26.

T-TN

, Jung

T-P

and Lin

C-T.

Retrosplenial segregation reflects the navigation load during ambulatory movement. IEEE Trans Neural Syst Rehabil Eng 2021; 29: 488–496.

27.

Desikan

, Ségonne

, Fischl

, . An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage 2006; 31: 968–980.

28.

Klein

and Tourville

101 Labeled brain images and a consistent human cortical labeling protocol. Front Neurosci 2012; 6: 171.

29.

Katayama

and Polich

Stimulus context determines P3a and P3b. Psychophysiology 1998; 35: 23–33.

30.

Wronka

, Kaiser

and Coenen

AMLL.

Neural generators of the auditory evoked potential components P3a and P3b. Acta Neurobiol Exp (Wars) 2012; 72: 51–64.

31.

Nichols

and Holmes

AP.

Nonparametric permutation tests for functional neuroimaging: A primer with examples. Hum Brain Mapp 2002; 15: 1–25.

32.

Downes

, Kirkham

, Telfer

, . Altered neurophysiological processing of auditory attention in preschool children with sickle cell disease. J Pediatr Psychol 2018; 43: 856–869.

33.

Youssry

, Shennawy

, Salam

, . Pattern of auditory and cognitive impairment in children with sickle cell disease: Single center experience. Pediatr Sci J 2023; 3: 1-10.

34.

Kirk

, Haynes

, Palasis

, . Regionally specific cortical thinning in children with sickle cell disease. Cereb Cortex 2009; 19: 1549–1556.

35.

Liddle

, Kiehl

and Smith

AM.

Event-related fMRI study of response inhibition. Hum Brain Mapp 2001; 12: 100–109.

36.

Mulert

, Pogarell

, Juckel

, . The neural basis of the P300 potential: Focus on the time-course of the underlying cortical generators. Eur Arch Psychiatry Clin Neurosci 2004; 254: 190–198.

37.

Utevsky

, Smith

and Huettel

SA.

Precuneus is a functional core of the default-mode network. J Neurosci Off J Soc Neurosci 2014; 34: 932–940.

38.

Kawadler

, Clayden

, Kirkham

, . Subcortical and cerebellar volumetric deficits in paediatric sickle cell anaemia. Br J Haematol 2013; 163: 373–376.